[Openmp-commits] [openmp] [openmp][amdgpu] Add DynamicLdsSize to AMDGPUImplicitArgsTy (PR #65325)

[via Openmp-commits]

https://github.com/skc7 created https://github.com/llvm/llvm-project/pull/65325:

Add DynamicLdsSize to AMDGPUImplicitArgsTy struct and fill the dynamic LDS size before kernel launch.

>From fad69d8c48c9c694abce9cbb75ee73df77696172 Mon Sep 17 00:00:00 2001
From: skc7 <Krishna.Sankisa at amd.com>
Date: Tue, 5 Sep 2023 14:46:05 +0530
Subject: [PATCH] [openmp][amdgpu] Add DynamicLdsSize to AMDGPUImplicitArgsTy

---
 openmp/libomptarget/plugins-nextgen/amdgpu/src/rtl.cpp         | 1 +
 .../libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h   | 3 ++-
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/openmp/libomptarget/plugins-nextgen/amdgpu/src/rtl.cpp b/openmp/libomptarget/plugins-nextgen/amdgpu/src/rtl.cpp
index 494fc66292e5bff..2003ce1013813e5 100644
--- a/openmp/libomptarget/plugins-nextgen/amdgpu/src/rtl.cpp
+++ b/openmp/libomptarget/plugins-nextgen/amdgpu/src/rtl.cpp
@@ -3010,6 +3010,7 @@ Error AMDGPUKernelTy::launchImpl(GenericDeviceTy &GenericDevice,
     ImplArgs->GroupSizeY = 1;
     ImplArgs->GroupSizeZ = 1;
     ImplArgs->GridDims = 1;
+    ImplArgs->DynamicLdsSize = KernelArgs.DynCGroupMem;
   }
 
   // Push the kernel launch into the stream.
diff --git a/openmp/libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h b/openmp/libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h
index b39545ab7d02ba2..d9828f0aa37a912 100644
--- a/openmp/libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h
+++ b/openmp/libomptarget/plugins-nextgen/amdgpu/utils/UtilitiesRTL.h
@@ -43,7 +43,8 @@ struct AMDGPUImplicitArgsTy {
   uint16_t GroupSizeZ;
   uint8_t Unused0[46]; // 46 byte offset.
   uint16_t GridDims;
-  uint8_t Unused1[190]; // 190 byte offset.
+  uint32_t DynamicLdsSize;
+  uint8_t Unused1[186]; // 186 byte offset.
 };
 
 // Dummy struct for COV4 implicitargs.